In-person, human verification of documents is relatively straight-forward. An individual provides a document (e.g., an identification (ID) or a credit card) to a third-party, and the third-party inspects (e.g., visually and tactilely) and verifies the document. However, such verification is subject to human error and malfeasance, as well as the cost of paying, training, and supervising human inspectors. Accordingly, the traditional observations may be inefficient and improved automatic systems and method for document verification are desirable.
Moreover, such traditional verification is not possible using remote systems. Meanwhile, the rise of internet transactions and smart phones has increased a need for remote verification of documents. For instance, to prove a purchaser owns a credit card used for a transaction, the user enters the card number, billing address, and card verification value (CVV). However, this information could be discovered and relayed by third parties and used to conduct a fraudulent purchase. Therefore, the use of a single type of verification (i.e., something you know) is insufficient to verify documents remotely. Thus, certain related art solutions require the provision of a photo of the document for verification verified. Continuing with the transaction example, not only must a purchaser provide the card number, billing address, and card verification value (CVV), the purchaser must also provide an image of their credit card. A transaction server validates the card for use in the transaction by verifying the provided information and by using image analysis on the image of the credit card. However, both person- and computer-based image analysis may be fooled by an image of copy a card or other document (e.g., a photocopy). For example, it may be difficult for a computer to differentiate between a picture of a credit card and a picture of a picture of the credit card.
Accordingly, there is a need for improved verification that may work either remotely or automatically.